lmao
>>107817231>in one experimentWhat about in two experiment?
>>107817231>(((in one experiment)))Ah yes repeat until you achieve the desired results.
>>107817231sometimes adequately explaining the required solution is more work than implementing the solution probably hurts experienced engineers the most
>>107817255>I don't know, AI did it like that, you can ask it, but the problem is solved, do you want the problem to be solved or do you want to be entertained with explanations you fucking cunt? approve the pull request, our users want shit to work and it's more important than your desire to know everything as if you need to know itproblem solved
>>107817231You use the agent in a way that it can automatically document and reuse work.If run 1 takes 2 hours, run 2 on the same task should take 2 mins.
>>107817231>thinking ai will make life easierwho wants to tell him?
>>107817231desu it works well for heaps of boilerplate such as GUI crapas much as I dislike aislop, I can't be bothered to write 500 LOC just to draw some sliders and buttons
Fake news
>>107817296Whole 2 min to copy and paste, sasuga
1. Retards tends to trust in those sloppers more than in humans (because normies use their retarded mindgames to evaluate others, that doesn't works with the slopper)2. retards under no circumstance will admit they used the slopper for the task and will not describe what parts are probably compromised with nil quality inputsThe end result is structural rot that can't be detected until it degrades everything to the point of being nonfunctional. Normies and their sloppers are 2 entry points to destroy supply chains (the end goal of sloppers is being an irreplaceable cog in that, even if originally they weren't needed)
>>107817372trv nk
>>107817382>6 GORILLION TREES DIED FOR THAT SLOP AND 7 BILLION INDIANS NOW CAN'T AFFORD RAM!
>>107817372nuclear-tier cope
>>107817231>>107817372This is a pretty bad headline. According to the study, it slowed them down across the board. The study seems to be a few months old but OP's article is only a few days old.>16 developers with moderate AI experience complete 246 tasks in mature projects on which they have an average of 5 years of prior experience. Each task is randomly assigned to allow or disallow usage of early 2025 AI tools. When AI tools are allowed, developers primarily use Cursor Pro, a popular code editor, and Claude 3.5/3.7 Sonnet. Before starting tasks, developers forecast that allowing AI will reduce completion time by 24%. After completing the study, developers estimate that allowing AI reduced completion time by 20%. Surprisingly, we find that allowing AI actually increases completion time by 19%--AI tooling slowed developers down.https://arxiv.org/abs/2507.09089
>>107817449The study literally boils down to AI tools not trained on your 1.1M+ LOC codebase aren't helping developers that have spent 5 years+ (10 really) on that codebase and it decreases productivity more due time taken for prompting and waiting on the AI generation than checking the output.That headline is fucking atrocious.
>>107817553>Cursor Pro, a popular code editor, and Claude 3.5/3.7 Sonnetbut that's what claude is
>>107817255sometimes adequately explaining the solution is the hardest part needed to fully think through the solution. i fins ai useful because its basically a rubber suck that can talk back and help me figure out what i actually need
>>107817231>subhuman silicon valley retards assumed that an evil ZOG botnet scambot would make things better, but it actually made them and everyone else who uses it slower and dumber >a few global bankers and billionaires profited though :)whoda thunk it
>>107818140They're just repeating what they read in ***media***, aka the scammers spokemen
>>107817241>>107817246why are AIbros so antisemitic
>>107817231My experience has been that it can complete tasks that would have taken me hours with a few minutes of prompting and another few minutes watching youtube while it runs. Some follow-ups and refactoring are often needed but it's just not comparable.I do find the resulting code quality is mildly to moderately worse than what I would have handwritten. For personal pet projects and open source contributions that might be a problem. For work I don't really care, it's good enough and meets the requirements and I can always refactor it later (probably with AI) if I feel like it.Also there's quite a big difference between Sonnet 3.x and Opus 4.5.
>>107817231The tasks took 20 minutes longer because AI is doing it RIGHT. Your average zoomy and shitlennial couldn't code a phone app. Have you not paid attention to the current state of software and how buggy and shit modern programs are? THAT is because zooms and shitlennials spent more time in college studying gender studies than actual mode. AND, most of them have zero passion for it, they only follow that path because they're lazy as fuck and want to work from home and take 30 breaks a day.Again AI DID TOOK LONGER BECAUSE IT HAD TO FIX THE SHIT STORM THE ZOOM OR SHITLENNIAL MADE.
>>107817231AIJeets get the rope
>>107818245>muh millennialscope harder you Gen Xir uncyou parented those zoomers
>>107818400>shitlinnial is PISSEDwhen you're done shitting and pissing yourself, calm down and take a hard hard look at worth useless and incompetent your generation is.
>>107818245>because AI is doing it RIGHTlook I don't mind knowingly taking the bait occasionally just for shits n giggles but you gotta meet me halfway here
>>107817246lol seething clanker cant keep up with junior codemonkeys
>>107818372I'll do one better.
>>107817553what a retarded cope. you are an indian AI shill.like everyone has been telling you that actually writes code, it doesn't work well on real projects.saars please redeeming the claude AI code very beautifal for gorgious good looks
